Measuring the load on picks in mining operations poses a significant challenge, which hinders the optimization of cutting mechanism design. To address this issue, a unified load measurement model was initially developed using angular transformation. Subsequently, the measuring-point positions were determined based on the principles of large strain and ease of measurement, using Finite Element Analysis. By analyzing the strain law of the measuring-point positions, a sensor layout scheme was designed, which led to the development of a triaxial load cell. Utilizing the piecewise least squares method and BP neural network, the calibration and decoupling were performed. The findings indicated that the BP neural network effectively reduced the error. Finally, cutting experiments were conducted. The error was less than 8 %, effectively identifying the load variation characteristics when cutting materials with different hardness. The results demonstrate that the designed triaxial load cell can accurately and reliably capture cutting loads.